Cancer is now the number one cause of death in North America. Malignant tumors of the central nervous system (CNS) are the third leading cause of cancer-related deaths in adolescents and adults between the ages of 15 and 34, and in children, brain tumors are the leading cause of cancer death. Furthermore, the two-year survival rate for patients with glioblastoma multiforme (GBM), a high-grade glioma (HGG), grade IV, is less than 20% (Davis et al. (1998) J. Neurosurg. 88:1-10), and there has been a steady increase in the incidence of brain cancers during the last 20 years (“Reports from the front” (1995) Science 267:1414). Almost any cancer can metastasize to the CNS (Olson et al. (1974) Arch. Neurol. 30:122-136).
A common approach to the treatment of malignant gliomas involves surgery (Berger (1994) Sem. Oncol. 21:172-185), radiation therapy (Gunderson & Tepper, Eds. (2000) Clinical Radiation Oncology (Churchill-Livingstone, Philadelphia), pp 314-35), and various chemotherapeutic regimens (Lesser & Grossman (1994) Sem. Oncol. 21:220-235), but neither single nor multimodal treatments are curative. At present, treatment is implemented to improve or sustain neurological function of the patient, to diminish the size of the tumor growing intracranially, and to lengthen intervals between treatments. Thus, new and molecular-specific methods of HGG treatment are urgently needed.
The transmembrane protein EphA2 is overexpressed and an attractive molecular target in glioblastoma multiforme (GBM), the most common primary brain tumor, which has a high incidence of recurrence and dim prognosis (Wykosky et al. (2005) Mol. Cancer. Res. 3:541-551). Only 3% of patients survive five years, with a median survival of approximately 14 months (Davis et al. (1998) J. Neurosurg. 88:1-10; Stupp et al. (2005) N. Engl. J. Med. 352:987-996). Interestingly, ephrinA1, a ligand for EphA2, is virtually not detectable in GBM cell lines, and is present at low levels in the majority of specimens despite the abundant overexpression of the receptor (Wykosky et al. (2005) Mol. Cancer. Res. 3:541-551). A similar pattern of differential EphA2/ephrinA1 expression has been reported in breast cancer (Macrae et al. (2005) Cancer Cell 8:111-118).
The ephrins comprise a family of protein ligands for the Eph receptor tyrosine kinases, and are unique among ligands for receptor tyrosine kinases in that they have been described as GPI-linked (ephrinA) or transmembrane (ephrinB) cell surface proteins rather than soluble factors (Davis et al. (1994) Science 266:816-819). Hence, it has been widely accepted that endogenous activation of Eph receptors by their ephrin ligands requires stable cell-cell contact and/or clustering by ephrin ligands (Shao et al. (1995) J. Biol. Chem. 270:5636-5641; Stein et al. (1998) Genes Dev. 12:667-678).
The physiological role of ephrins and Eph receptors lies primarily in the formation and organization of the vasculature (McBride et al. (1998) Mech. Dev. 77:201-204; Wang et al. (1998) Cell 93:741-753) and the patterning of topographic maps in the developing retinotectal and central nervous systems, where Eph signaling transmits cues for axon guidance (Drescher et al. (1995) Cell 82:359-370; Nakamoto et al. (1996) Cell 86:755-766; Knoll et al. (2002) Trends Neurosci. 25:145-149; Rashid et al. (2005) Neuron 47:57-69). Furthermore, some Eph receptors, including EphA2 and EphB2 (Nakada et al. (2004) Cancer Res. 64:3179-3185), have been implicated in tumorigenesis. Specifically, the EphA2 receptor is overexpressed and functionally important in cancers of the breast (Zelinski et al. (2001) Cancer Res. 61:2301-2306; Wu et al. (2004) Pathol. Oncol. Res. 10:26-33), prostate (Walker-Daniels et al. (1999) Prostate 41:275-280; Zeng et al. (2003) Am. J. Pathol. 163:2271-2276), brain (Wykosky et al. (2005) Mol. Cancer. Res. 3:541-551; Hatano et al. (2005) Neoplasia 7:717-722), and ovary (Thaker et al. (2004) Clin. Cancer Res. 10:5145-5150).
Due to the absence of the receptor in normal brain, opportunities exist for EphA2-targeted therapies based on ephrinA1, similar in high GBM cell molecular-specificity to what has been shown previously for another glioma-associated antigen, IL-13 receptor alpha-2 (Debinski et al. (1998) Nat. Biotechnol. 16:449-453). Interestingly, the ligand for EphA2, ephrinA1, is on average expressed (i) at lower levels when the receptor is elevated, and (ii) at higher levels when the receptor is low. Hence, it was hypothesized that ephrinA1 is a tumor-suppressing protein in several solid tumors. Furthermore, a soluble recombinant homodimer, ephrinA1-Fc, activates EphA2 in GBM and other tumor cells, resulting in alteration of their malignant features. However, the prevailing notion has been that ephrinA1 must be anchored to the plasma membrane and form a oligodimer in order to activate EphA2 in malignancy.
Previously, the tumor-suppressing potential of ephrinA1 has been demonstrated using, e.g., a recombinant, covalently-linked homodimeric ephrinA1-Fc. This protein activates and reverses the oncogenic properties of EphA2 in tumor cells in culture, causing receptor down-regulation, cell morphology changes, suppression of integrin function, and negative regulation of invasion, migration, and anchorage-independent growth (Wykosky et al. (2005) Mol. Cancer. Res. 3:541-551); (Miao et al. (2000) Nat. Cell Biol. 2:62-69; Carter (2002) Nat. Cell Biol. 4:565-573; Walker-Daniels et al. (2002) Mol. Cancer. Res. 1:79-87; Duxbury et al. (2004) Biochem. Biophys. Res. Commun. 320:1096-1102). The known physiological roles of ephrinA1, also thought to be mediated by a membrane-anchored form of the ligand and dependent on cell-cell contact, include induction of cell repulsion and growth cone collapse during central nervous system development (Marquardt et al. (2005) Cell 121:127-139). In addition, ephrinA1 has been shown to be expressed in the developing vasculature during embryogenesis (McBride et al. (1998) Mech. Dev. 77:201-204), induces endothelial cell migration (Pandey et al. (1995) Science 268:567-569) and the formation of capillary-like structures in vitro (Daniel et al. (1996) Kidney Int. Suppl 57:S73-S81), and plays a role in angiogenesis and neovascularization in vivo (Cheng et al. (2002) Mol. Cancer. Res. 1:2-11).
EphrinA1 was originally isolated as a TNF-α-inducible, immediate-early response gene product from normal human umbilical vein endothelial cells (Holzman et al. (1990) Mol. Cell Biol. 10:5830-5838). The full-length mature protein is composed of 187 amino acids with a molecular weight of 22 kDa. The C-terminus of ephrinA1, due to its hydrophobic nature interrupted by several charged amino acids extending to the extreme C-terminal end, has high structural similarity to GPI-linked, membrane-anchored proteins (Ferguson et al. (1988) Annu. Rev. Biochem. 57:285-320). Interestingly, however, the original study did not provide evidence for its presence on the surface of normal endothelial cells. It was only later demonstrated that the ligand may exist as a GPI-linked protein in ephrinA1-transfected human embryonic kidney and breast cancer cells by its release into the media upon proteolytic cleavage with phosphatidylinositol-specific phospholipase C (PI-PLC) (Shao et al. (1995) J. Biol. Chem. 270:5636-5641). Additional indirect support for the membrane-anchored presence of the ligand was in the finding that soluble ephrinA1 could activate the EphA5 receptor only when clustered by antibodies against C-terminal epitope tags (Davis et al. (1994) Science 266:816-819).
These observations, coupled with structural studies on active Eph/ephrin complexes (Himanen et al. (2001) Nature 414:933-938; Toth et al. (2001) Dev. Cell 1:83-92), gave rise to the notion that clustering of ephrins is a process necessary for Eph receptor activation that can be accomplished in a number of ways: via membrane attachment (Davis et al. (1994) Science 266:816-819; Shao et al. (1995) J. Biol. Chem. 270:5636-5641), antibody-mediated clustering (Davis et al. (1994) Science 266:816-819), or the formation of soluble homodimers through disulfide bonding of an IgG-Fc conjugate (Stein et al. (1998) Genes Dev. 12:667-678). Hence, membrane-bound ephrinA1 is considered the endogenous, functional form of the ligand (Beckmann et al. (1994) EMBO J. 13:3757-3762; Xu et al. (1997) J. Mol. Med. 75:576-586; Kullander et al. (2002) Nat. Rev. Mol. Cell. Biol. 3:475-486; Pasquale (2005) Nat. Rev. Mol. Cell Biol. 6:462-475). The majority of studies on the role of ephrinA1 and other ephrinA's, both in physiology and in tumorigenesis, employ the Fc-conjugated dimeric forms of the ligand, often pre-clustered by the addition of IgG (Davis et al. (1994) Science 266:816-819; Daniel et al. (1996) Kidney Int. Suppl. 57:S73-S81).
There remains a need to simplify and more efficiently utilize ephrinA1, e.g., for its anti-tumor activity.